Yes, A Robot Could Do Your Job

Here's a game to play over dinner. One person names a profession that they believe can't be taken over by a machine, and another person has to make a case why it's not so future-proof.

Picture: Peyri Herrera

We played this game on an episode of SBS's Insight on the topic of the future of robots and artificial intelligence.

The first profession suggested was musician. An argument often put forwards against artificial intelligence (AI) is that computers can't be creative. But there are plenty of examples to counter this argument. For instance, computers can take plain sheet music and turn it into an expressive jazz performance, as my colleague Ramon Lopez de Mantaras has shown.

So, jazz musicians watch out. Your jobs might not be safe from robot incursion.

The next option was police officer. It's often said that computers can't or won't behave ethically. Unfortunately, Hollywood has already painted a very dystopian picture here in movies like Robocop and Terminator. And, as the current UN campaign to ban autonomous weapons demonstrates, we could easily end up there if we aren't careful.

The third profession put forward was human resources. Naturally, this came from an HR consultant worried for her future job prospects. However, the bureaucratic side of HR is already easily automated. Indeed, we spend much of our lives on the phone already talking to machines. Can I speak to a real person, please?

On the other hand, the more human-facing side of HR is likely to be harder to automate. But as we argue in the next answer, it's not clear that this will be impossible.

The fourth challenge was psychiatrist. Again, the human-facing nature of this would seem to offer significant resistance to automation. Nevertheless, there's an interesting historical precedent.

A well-known computer program called Eliza was the very first chatterbot. It unintentionally passed itself off as a real Rogerian psychotherapist.

Eliza was not very smart. Indeed, the program's author, Joseph Weizenbaum, meant it more as parody than as therapist. However, his secretary famously asked to be left alone so she could talk in private to the chatterbot.

So, shrinks watch out. Your jobs might not be safe.

The final challenge was Prime Minister.

On the one hand, this is a good answer, as one assumes there's little routine to being Prime Minister but a lot of tough high level decision making that would be tough for a machine to handle. On the other hand, it's a poor winner of our little game. It may be the only job in the whole country that's safe from robots.

In one final, beautiful irony, this episode of Insight has the robots up on the stage. We, the supposed expert commentators were in the audience. So, even TV pundits should watch out. Your jobs might not be safe too.

Net effects

What this discussion highlights is that the middle classes are likely to be increasingly squeezed by machine labour. Professions that we used to think were quite safe -- like doctor, lawyer or accountant -- will be increasingly automated.

Whenever technology takes away jobs, it tends to make new jobs and industries elsewhere. For example, printing removed the need for scribes but created the vast publishing industry in its stead. And publishing went on to create many other jobs in the industries that grew out of all the knowledge passed on in printed material.

More recently, computers have taken away many traditional jobs in the printing industry, like type setters. But we now see many new jobs in areas like self-publishing and web design.

Economists continue to argue over the net effects of technology. Does technology create more economic activity so we are all better off? Or does it put more people out of work, concentrating wealth in the hands of the few?

One thing seems sure. It requires us to adapt. And for this, we need an educated, high tech workforce. This brings the conversation back to higher education and the stalled reforms that now trouble this sector in Australia.

If there is one policy we need to get right, to future-proof Australia against machines and other disruptions, I would argue, this is it.The Conversation

Toby Walsh is Professor, Research Group Leader, Optimisation Research Group at NICTA.

This article was originally published on The Conversation. Read the original article.


Comments

    Regarding replacing HR - what's the point of HR if you have replaced all your human employees with robots?

    HR is totally able to be made obsolete by robots.

    A robot could not be a religious priest, or a theologian. Robots will generally come from a logical point of view, and the self-contradictory nature of the major religions would be a roadblock that they simply wouldn’t be able to pass. How would the robot know when to make a leap of faith, or what specific old doctrines should be ignored (no shellfish everyone), while other seemingly capricious commandments must be slavishly followed (no homosexual relationships please).
    Religion is such a manmade affair that not only would you have to program the artificial intelligence to understand the stuff, you would have to actually make it want to believe in an afterlife, magical god, and so on.
    Can’t see it happening.
    But then again, we are just biological machines ourselves, so whatever…

      I can't tell if you're for or against your own argument. The things you've pointed out aren't requirements, just beliefs (from an evolutionary and survival standpoint), and some minor roles/ideologies associated with them. Elaborate?

    Graphic Designer. No robot could deal with contradictory, illogical clients. How would a robot react to "make it pop"?

      Actually thinking about it, dealing with clients on a design project would easily could easily drive a robot to go all 'Skynet'.

    The problem with robots is that they're made and programmed by people, which means there's always bugs, faults and poorly handled unexpected conditions. It is possible to have close to perfect systems, however the level of investment to implement such solutions can easily exceed the cost of having the job performed by a person in the first place.

    This comes from someone with hands on experience in both enterprise IT, and large scale highly automated manufacturing (which ends up needing constant human intervention on a very frequent basis to keep running).

Join the discussion!